KOZO KEIKAKU ENGINEERING Inc.
RiverCast: A Cloud System That Predicts River Levels 15 Hours Ahead
| Updated | August 4, 2022 |
|---|---|
| Publication date | December 25, 2020 |
| Sector | Natural Disasters / Coastal Areas |
Company Overview
KOZO KEIKAKU ENGINEERING Inc., which began in 1956 with structural design work for buildings, provides analysis and simulation of natural phenomena (earthquakes, tsunamis, wind, etc.) surrounding man-made structures, software development in the field of information and communications, sales and customization of CAD/CAE software for the manufacturing sector, and consulting in the field of supporting human decision making to help solve problems that the society and communities are facing.
Climate Change Impacts
The role of flood forecasting systems is becoming increasingly important as climate change causes more intense rainfall and flood damage. However, the use of conventional methods has not been widespread except for major rivers because they require a substantial amount of time and cost to implement. Climate change may also make it difficult to forecast floods of unprecedented magnitude.
Adaptation Initiatives
We have developed a flood forecasting system RiverCast in collaboration with the University of Tokyo by applying the most advanced mathematical engineering. This system has the following features.
- The system can be deployed faster and with less cost compared to existing methods, and can also be applied to small to medium-sized rivers.
- Using state-of-the-art mathematical engineering technology, the system can predict water levels with high accuracy with little training data, and is capable of predicting floods of unprecedented scale.
- Water level prediction is strongly affected by errors in weather forecasts, but RiverCast provides probabilistic water level prediction that takes rainfall errors into account, and can display the confidence interval of the prediction and the probability of exceeding the standard water level, just like the probability circle and rainfall probability of a typhoon.
- As a cloud service, the system can be easily viewed from a computer or smartphone, with the function to notify the user by e-mail 24 hours a day, 365 days a year, at a specified time or when the water level is predicted to rise.
The effectiveness of this system has been verified by introducing the technology and conducting demonstration tests at more than 45 sites in 13 municipalities nationwide, including Kawasaki City, Yokohama City, and Osaka Prefecture.
Effects / Expected Benefits
This system provides positive prediction results compared to existing methods, and its effectiveness in actual flood control activities has been confirmed.
We are currently developing projects using this system, and many private companies and municipalities have decided to deploy the system. In addition to evacuation decisions and flood prevention activities, the system is used for facility management around rivers and for safety management at construction sites.
